Meta-Analysis of Genome-Wide Association Studies Identifies Eight New Loci for Type 2 Diabetes in East Asians
Total Page:16
File Type:pdf, Size:1020Kb
LETTERS Meta-analysis of genome-wide association studies identifies eight new loci for type 2 diabetes in east Asians Yoon Shin Cho1,46, Chien-Hsiun Chen2,3,46, Cheng Hu4,46, Jirong Long5,46, Rick Twee Hee Ong6,46, Xueling Sim7,46, Fumihiko Takeuchi8,46, Ying Wu9,46, Min Jin Go1,46, Toshimasa Yamauchi10,46, Yi-Cheng Chang11,46, Soo Heon Kwak12,46, Ronald C W Ma13,46, Ken Yamamoto14,46, Linda S Adair15, Tin Aung16,17, Qiuyin Cai5, Li-Ching Chang2, Yuan-Tsong Chen2, Yutang Gao18, Frank B Hu19, Hyung-Lae Kim1,20, Sangsoo Kim21, Young Jin Kim1, Jeannette Jen-Mai Lee22, Nanette R Lee23, Yun Li9,24, Jian Jun Liu25, Wei Lu26, Jiro Nakamura27, Eitaro Nakashima27,28, Daniel Peng-Keat Ng22, Wan Ting Tay16, Fuu-Jen Tsai3, Tien Yin Wong16,17,29, Mitsuhiro Yokota30, Wei Zheng5, Rong Zhang4, Congrong Wang4, Wing Yee So13, Keizo Ohnaka31, Hiroshi Ikegami32, Kazuo Hara10, Young Min Cho12, Nam H Cho33, Tien-Jyun Chang11, Yuqian Bao4, Åsa K Hedman34, Andrew P Morris34, Mark I McCarthy34,35, DIAGRAM Consortium36, MuTHER Consortium36, Ryoichi Takayanagi37,47, Kyong Soo Park12,38,47, Weiping Jia4,47, Lee-Ming Chuang11,39,47, Juliana C N Chan13,47, Shiro Maeda39,47, Takashi Kadowaki10,47, Jong-Young Lee1,47, Jer-Yuarn Wu2,3,47, Yik Ying Teo6,7,22,25,41,47, E Shyong Tai22,42,43,47, Xiao Ou Shu5,47, Karen L Mohlke9,47, Norihiro Kato8,47, Bok-Ghee Han1,47 & Mark Seielstad25,44,45,47 We conducted a three-stage genetic study to identify have been identified for T2D10,11. Despite these advances toward a susceptibility loci for type 2 diabetes (T2D) in east Asian better understanding of the genetic basis of T2D, its heritability has populations. We followed our stage 1 meta-analysis of eight not been fully explained12. In addition, most of the T2D loci were T2D genome-wide association studies (6,952 cases with T2D detected initially in population samples of European origin, with the Nature America, Inc. All rights reserved. All rights Inc. America, Nature 2 and 11,865 controls) with a stage 2 in silico replication exceptions of KCNQ1, UBE2E2 and C2CD4A-C2CD4B, which were analysis (5,843 cases and 4,574 controls) and a stage 3 de novo first identified in studies of east Asian populations13–15. Additional © 201 replication analysis (12,284 cases and 13,172 controls). efforts involving east Asian populations identified variants associated The combined analysis identified eight new T2D loci reaching with T2D at the SPRY2, PTPRD and SRR loci5,16,17. However, these genome-wide significance, which mapped in or near GLIS3, associations need more validation from additional studies of east PEPD, FITM2-R3HDML-HNF4A, KCNK16, MAEA, GCC1-PAX4, Asians as well as in studies of other populations. A large meta-analysis PSMD6 and ZFAND3. GLIS3, which is involved in pancreatic in east Asians would be expected to identify new genetic associations beta cell development and insulin gene expression1,2, is known and provide insights into T2D pathogenesis. In addition to differences for its association with fasting glucose levels3,4. The evidence of in the allele frequencies between east Asians and Europeans, which an association with T2D for PEPD5 and HNF4A6,7 has been may affect the power to detect associations in these populations, T2D shown in previous studies. KCNK16 may regulate glucose- epidemiology also differs considerably between European populations dependent insulin secretion in the pancreas. These findings, and east Asian populations. In east Asians, the rates of diabetes are derived from an east Asian population, provide new often higher at lower average body mass indices (BMIs)18, suggesting perspectives on the etiology of T2D. that some different pathways may be involved in pathogenesis of T2D in east Asians and Europeans. T2D is a major public health problem with a rapidly rising global To discover new T2D loci, we conducted a three-stage association prevalence8. The development of T2D is influenced by diverse factors, study in individuals of east Asian descent (Supplementary Fig. 1). and decades of epidemiological studies have linked obesity, hyper- We performed the stage 1 meta-analysis by combining eight T2D tension and dyslipidemia with the risk of T2D9. It is also known that genome-wide association studies (GWAS) participating in the Asian T2D has considerable heritability. Within only the last 3 years, genetic Genetic Epidemiology Network (AGEN) consortium (6,952 cases studies have produced a rapidly lengthening list of loci harboring and 11,865 controls) with association data for 2,626,356 imputed and disease-predisposing variations10. To date, genetic variants at 45 loci genotyped autosomal SNPs, and we used the inverse-variance method A full list of affilations are at the end of article. Received 12 April; accepted 2 November; published online 11 December 2011; doi:10.1038/ng.1019 NATURE GENETICS VOLUME 44 | NUMBER 1 | JANUARY 2012 67 LETTERS for fixed effects for the statistical analyses (Supplementary Table 1). near FITM2-R3HDML-HNF4A identified in our study (rs6017317) is All imputed and genotyped SNPs (minor allele frequency (MAF) > not in strong LD with the HNF4A P2 promoter SNPs (r2 = 0.23–0.25, 0.01) passed quality control filters in each of the eight stage 1 datasets D′ = 0.50–0.54 between rs6017317 and rs1884613 or rs2144908 based prior to conducting the meta-analysis (Supplementary Table 2). The on phase 2 CHB/JPT HapMap data), indicating that rs6017317 is a genomic control inflation factor (λ) for the meta-analysis was 1.046 new T2D signal in the 20q13.12 region where HNF4A resides. (and was less than 1.062 for each of the individual studies), indicat- The other five loci reaching genome-wide significance in our ing that the results seen in stage 1 were probably not the result of study have not previously been reported in the context of any meta- population stratification (Supplementary Fig. 2). Individuals from bolic traits, including the loci mapped in or near KCNK16, MAEA, each component study that participated in stage 1 mainly clustered GCC1-PAX4, PSMD6 and ZFAND3. KCNK16, which is expressed together with the samples from the CHB/JPT HapMap population predominantly in the pancreas, encodes a potassium channel pro- in the principal component analysis plot (Supplementary Fig. 3), tein containing two pore-forming P domains23. In pancreatic β cells, further showing the similarity in ethnicity between the stage 1 samples. potassium channels that are inhibited by ATP regulate glucose- Most signals showing strong evidence for T2D associations were dependent insulin secretion. Among the variants in strong LD with in known T2D genes (Fig. 1). Stage 1 P values, odds ratios (ORs) the signal reaching genome-wide significance in KCNK16 (rs1535500) and average risk allele frequencies for 45 previously reported T2D- is rs11756091 (r2 = 0.977, D′ = 1.0 based on phase 2 CHB/JPT HapMap associated SNPs are listed in Supplementary Table 3. data), which encodes a substitution of proline to histidine in two After removing known T2D variants, we selected 297 SNPs from isoforms of KCNK16. This variant or others influencing KCNK16 independent loci from the stage 1 meta-analysis based on our arbi- may result in the defective regulation of potassium channel activity trary inclusion criteria for in silico follow-up replication: meta-analysis that contributes to the etiology of T2D24. MAEA encodes a protein P < 5 × 10−4 (based on the divergence between the observed and that has a role in erythroblast enucleation and in the development expected P values on the quantile-quantile plot; Supplementary of mature macrophages25. A gene-set analysis of the stage 1 P values Fig. 2), heterogeneity P > 0.01 and at least seven studies having been using GSA-SNP26 indicated that MAEA belongs to a group of genes included in the meta-analysis (Supplementary Table 4). We took a that previously showed significant association with T2D and includes total of 3,756 SNPs, including the 297 selected SNPs and their proxies IDE, which is located at a known T2D susceptibility locus27 (stage 1 (r2 > 0.8 based on phase 2 CHB/JPT HapMap data), forward to stage 2 P = 1.41 × 10−7 for rs6583826 at the IDE locus in this study). The (in silico replication) in three independent GWAS (5,843 cases and GRIP-domain–containing protein that is encoded by GCC1 might 4,574 controls). After a meta-analysis that combined stage 1 and have a role in the organization of the trans-Golgi network, which is 2 data for 3,756 SNPs, we selected the 19 SNPs that showed the involved in membrane transport28. PAX4, which is only 30 kb away most compelling evidence for association (stage 1 and 2 combined from GCC1, is a good candidate for T2D given its involvement in P < 10−5) (Supplementary Table 5) for stage 3 de novo genotyping in pancreatic islet development. PAX4 was recently implicated in a up to 12,284 cases and 13,172 controls recruited from five independ- Japanese individual with maturity onset diabetes of the young29. The ent studies (Supplementary Tables 1 and 2). This resulted in eight expression product of PSMD6, which acts as a regulatory subunit new T2D loci that reached genome-wide significance in the combined of the 26S proteasome, is probably involved in the ATP-dependent meta-analysis across all three stages (Table 1 and Fig. 2). degradation of ubiquitinated proteins30. Although the function of Three of these T2D-associated loci were previously associated with ZFAND3 has not been fully elucidated, it is noteworthy that a member Nature America, Inc. All rights reserved. All rights Inc. America, Nature metabolic traits or related diseases or were suggestively associated of the same gene family, ZFAND6, is present along with FAH at a 2 with T2D.